泵控缸电液位置伺服系统的神经网络模型参考自适应控制

Neural Network Model Reference Adaptive Control of Electro-hydraulic Position Servo System of Pump-controlled Cylinder

针对泵控缸电液位置伺服系统的跟踪控制问题,提出了神经网络模型参考自适应控制方法。泵控缸电液位置伺服系统由于其自身特性以及外界干扰因素的影响存在严重的非线性,因此,很难采用传统的控制方法来控制。为此,首先利用GA-BP算法离线辨识伺服系统的神经网络模型,得到网络参数的初值,然后利用改进的BP算法在线对网络参数进行微调,以得到较为准确的网络预测输出,从而为在线神经网络控制提供较准确的梯度信息。仿真结果表明,该方法能保证系统具有较快的响应速度和较高的控制精度,并具有较好的自适应性和鲁棒性。

A neural network model reference adaptive control approach was proposed for tracking control problems of the electrohydraulic position servo system of the pump-controlled cylinder. Since the system presents a severe nonlinearity due to its own characteristics and external interfering factors, it is difficult to get a satisficd control effect by a conventional method. In order to get exact forecasted output of network and gradient information used for the neural network control, the GA-BP was used to optimize the values of weights and thresholds off-line, and an optimized initial value was obtained. The modified BP algorithm was applied to find out an optimal values on-line. The simulation results show that the proposed approach guarantees the response speed and accuracy, which possesses a good self-adaptive and strong robust.